Parallel Disk Systems Research Articles

Coriolis effects on torque transmission of hydro-viscous film in parallel disks with imposed throughflow

The imposed throughflow into rotating parallel disk system is to decrease temperature rise of hydro-viscous film flow, but the induced Coriolis effects are usually underestimated on torque transmission. The analytical expressions of thin-film parameters are obtained based on Navier-Stokes equations with the method of approximate analytical iteration. The Coriolis force and the corresponding influence on torque transmission are analyzed by analytical calculation and experimental verification. The results show that the tangential Coriolis force serves as resistance to the viscous torque transmission and the Coriolis torque loss can lead to primary reduction of efficiency among all motion effects. Therefore, the Coriolis effects should be paid more attention for the practical designs and applications of rotating hydro-viscous parallel-disk systems with imposed throughflow.

Stegtorrent using De-Clustering of Data

Nowadays, For uploading and downloading of data using torrent people widely use internet. Existing system was less secure and time consuming. To avoid this problem stegtorrent using de-clustering of data is implemented. For file transfer service, StegTorrent is a new network steganographic method. To achieve availability and scalability of storage systems are important for existing information systems.One approach to achieving high availability of parallel disk systems is to replicate the data items on separate disk drives. Many applications require considerable space, which is increasing rapidly, and essential data kept in storage must be retained. Parallel storage configurations with multiple disk drives are used to satisfy these requirements. Because of the disk failure, when one copy is damaged then other can continue to be used. In the replication-based approach, it is also important to lower the cost of update operations. To ensure that the data mapping and requests are evenly divided among disks. In replicationbased approach, it takes the advantages of low cost data modification and data recovery.

MINT: A Reliability Modeling Frameworkfor Energy-Efficient Parallel Disk Systems

The Popular Disk Concentration (PDC) technique and the Massive Array of Idle Disks (MAID) technique are two effective energy conservation schemes for parallel disk systems. The goal of PDC and MAID is to skew I/O load toward a few disks so that other disks can be transitioned to low power states to conserve energy. I/O load skewing techniques like PDC and MAID inherently affect reliability of parallel disks, because disks storing popular data tend to have high failure rates than disks storing cold data. To study reliability impacts of energy-saving techniques on parallel disk systems, we develop a mathematical modeling framework called MINT. We first model the behaviors of parallel disks coupled with power management optimization policies. We make use of data access patterns as input parameters to estimate each disk's utilization and power-state transitions. Then, we derive each disk's reliability in terms of annual failure rate from the disk's utilization, age, operating temperature, and power-state transition frequency. Next, we calculate the reliability of PDC and MAID parallel disk systems in accordance with the annual failure rate of each disk in the systems. Finally, we use real-world trace to validate out MINT model. Validation result shows that the behaviors of PDC and MAID which are modeled by MINT have a similar trend as that in the real-world.

Central and Distributed GPU based Parallel Disk Systems for Data Intensive Applications

Parallel disk systems are capable of fulfilling rapidly increasing demands on both large storage capacity and high I/O performance. However, it is challenging to significantly increase disk I/O bandwidth for data-intensive workloads due to (1) reliability and instant processing of data requests under dynamic workload conditions, and (2) the optimum tradeoff between system scalability and data reliability in data-intensive systems. To increase computing performance and reduce power consumption, Graphics Processing Units (GPUs) will be used. As the architectures and data processing algorithms for GPU-based parallel disk systems are still in their infancy, this research will develop novel hardware and software architectures that include parallel GPU, flash disks, and disk arrays for data-intensive applications.

A Two-Pass Exact Algorithm for Selection on Parallel Disk Systems.

Numerous OLAP queries process selection operations of "top N", median, "top 5%", in data warehousing applications. Selection is a well-studied problem that has numerous applications in the management of data and databases since, typically, any complex data query can be reduced to a series of basic operations such as sorting and selection. The parallel selection has also become an important fundamental operation, especially after parallel databases were introduced. In this paper, we present a deterministic algorithm Recursive Sampling Selection (RSS) to solve the exact out-of-core selection problem, which we show needs no more than (2 + ε) passes (ε being a very small fraction). We have compared our RSS algorithm with two other algorithms in the literature, namely, the Deterministic Sampling Selection and QuickSelect on the Parallel Disks Systems. Our analysis shows that DSS is a (2 + ε)-pass algorithm when the total number of input elements N is a polynomial in the memory size M (i.e., N = Mc for some constant c). While, our proposed algorithm RSS runs in (2 + ε) passes without any assumptions. Experimental results indicate that both RSS and DSS outperform QuickSelect on the Parallel Disks Systems. Especially, the proposed algorithm RSS is more scalable and robust to handle big data when the input size is far greater than the core memory size, including the case of N ≫ Mc .

Strip-oriented asynchronous prefetching for parallel disk systems

Sequential prefetching schemes are widely employed in storage servers to mask disk latency and improve system throughput. However, existing schemes cannot benefit parallel disk systems as expected due to the fact that they ignore the distinct internal characteristics of the parallel disk system, in particular, data striping. Moreover, their aggressive prefetching pattern suffers from premature evictions and prolonged request latencies. In this paper, we propose a strip-oriented asynchronous prefetching (SoAP) technique, which is dedicated to the parallel disk system. It settles the above-mentioned problems by providing multiple novel features, e.g., enhanced prediction accuracy, adaptive prefetching strength, physical data layout awareness, and timely prefetching. To validate SoAP, we implement a prototype by modifying the software redundant arrays of inexpensive disks (RAID) under Linux. Experimental results demonstrate that SoAP can consistently offer improved average response time and throughput to the parallel disk system under non-random workloads compared with STEP, SP, ASP, and Linux-like SEQPs.

An adaptive energy-conserving strategy for parallel disk systems

Although various parallel disk systems have been developed to achieve high I/O performance and energy efficiency, most existing parallel disk systems lack an adaptive way to conserve energy in dynamically changing workload conditions. To solve this problem, we develop an adaptive energy-saving scheme or DCAPS in parallel disk systems. We show that adaptability in energy conservation can be achieved through the integration of a dynamic disk scheduling scheme and power management in parallel disk systems. DCAPS consists of a data partitioning mechanism, a response time estimator, and an adaptive energy-conserving mechanism. The Data partitioning mechanism allows DCAPS to adjust the parallelism degrees of write requests based on dynamic workload conditions. Apart from supporting the data partitioning mechanism, the response time estimator makes it possible for the adaptive energy-conserving mechanism to dynamically adjust voltage supply levels while guaranteeing desired response times. We conducted extensive experiments to quantitatively evaluate the performance of the proposed energy-conserving strategy. Experimental results consistently show that DCAPS significantly reduces energy consumption of parallel disk systems in a dynamic environment over the same disk systems without using DCAPS.

An Improved Compiler Directed Power Optimization for Disk Based Systems using Back Propagation Neural Networks

The performance of the parallel disk systems is highly affected by excessive power consumption. Hence, various researches are being done in the areas of energy optimization of such systems. Because of the extensive growth and development of computer applications, there has been a huge transformation in the disk subsystem, which mainly includes larger number of disks with higher storage capacities and rotational speeds. Hence, Disk power management has become an essential area of research in recent years as it utilizes very high power. An advanced compiler-directed disk power management approach is presented in this paper which exploits disk access approaches for minimizing energy consumption. This paper examines the various disk access techniques and chooses the optimal approach which would offer better overall performance of the disks via Back Propagation neural network technique which is trained using Modified Levenberg Morquat learning. The experimental evaluation shows that the proposed technique offers better power consumption than the traditional approaches.

PRE-BUD

A critical problem with parallel I/O systems is the fact that disks consume a significant amount of energy. To design economically attractive and environmentally friendly parallel I/O systems, we propose an energy-aware prefetching strategy (PRE-BUD) for parallel I/O systems with disk buffers. We introduce a new architecture that provides significant energy savings for parallel I/O systems using buffer disks while maintaining high performance. There are two buffer disk configurations: (1) adding an extra buffer disk to accommodate prefetched data, and (2) utilizing an existing disk as the buffer disk. PRE-BUD is not only able to reduce the number of power-state transitions, but also to increase the length and number of standby periods. As such, PRE-BUD conserves energy by keeping data disks in the standby state for increased periods of time. Compared with the first prefetching configuration, the second configuration lowers the capacity of the parallel disk system. However, the second configuration is more cost-effective and energy-efficient than the first one. Finally, we quantitatively compare PRE-BUD with both disk configurations against three existing strategies. Empirical results show that PRE-BUD is able to reduce energy dissipation in parallel disk systems by up to 50 percent when compared against a non-energy aware approach. Similarly, our strategy is capable of conserving up to 30 percent energy when compared to the dynamic power management technique.

Quality of security adaptation in parallel disk systems

In the past decade, parallel disk systems have been highly scalable and able to alleviate the problem of disk I/O bottleneck, thereby being widely used to support data-intensive applications. Although a variety of parallel disk systems were developed, most existing disk systems lack a means to adaptively control the quality of security for dynamically changing workloads. We address this gap in disk technology by designing, implementing, and evaluating a quality of security control framework for parallel disk systems, or ASPAD for short, that makes it possible for parallel disk systems to adapt to changing security requirements and workload conditions. The ASPAD framework comprises four major components, namely, a security service middleware, a dynamic data-partitioning mechanism, a response time estimator, and an adaptive security quality controller. The framework is conducive to adaptively and expeditiously determining security services for requests submitted to a parallel disk system in a way to improve security of the disk system while making an effort to guarantee desired response times of the requests. We conduct extensive experiments to quantitatively evaluate the performance of the proposed ASPAD framework. Empirical results show that ASPAD significantly improves the overall performance of parallel disk systems over the same disk systems without using the ASPAD framework.

A reliability model of energy-efficient parallel disk systems with data mirroring

In the last decade, parallel disk systems have increasingly become popular for data-intensive applications running on high performance computing platforms. Conservation of energy in parallel disk systems has a strong impact on the cost of cooling equipment and backup power-generation. This is because a significant amount of energy is consumed by parallel disks in high performance computing centres. Although a wide range of energy conservation techniques have been developed for disk systems, most energy saving schemes have adverse impacts on the reliability of parallel disk systems. To address this deficiency, we must focus on reliability analysis for energy-efficient parallel disk systems. In this paper, we make use of a Markov process to develop a quantitative reliability model for energy-efficient parallel disk systems using data mirroring. With the new model in place, a reliability analysis tool is developed to efficiently evaluate reliability of fault-tolerant parallel disk systems with two power modes. More importantly, the reliability model makes it possible to provide good trade-offs between energy efficiency and reliability in energy-efficient and fault-tolerant parallel disk systems.

Integrated prefetching and caching in single and parallel disk systems

We study integrated prefetching and caching in single and parallel disk systems. In the first part of the paper we investigate approximation algorithms for the single disk problem. There exist two very popular approximation algorithms called Aggressive and Conservative for minimizing the total elapsed time. We give a refined analysis of the Aggressive algorithm, improving the original analysis by Cao et al. We prove that our new bound is tight. Additionally we present a new family of prefetching and caching strategies and give algorithms that perform better than Aggressive and Conservative . In the second part of the paper we investigate the problem of minimizing stall time in parallel disk systems. We present a polynomial time algorithm for computing a prefetching/caching schedule whose stall time is bounded by that of an optimal solution. The schedule uses at most 2 ( D - 1 ) extra memory locations in cache. This is the first polynomial time algorithm that, using a small amount of extra resources, computes schedules whose stall times are bounded by that of optimal schedules not using extra resources. Our algorithm is based on the linear programming approach of [Journal of the ACM, 47 (2000) 969]. However, in order to achieve minimum stall times, we introduce the new concept of synchronized schedules in which fetches on the D disks are performed completely in parallel.

RAPID-Cache-A reliable and inexpensive write cache for high performance storage systems

Modern high performance disk systems make extensive use of nonvolatile RAM (NVRAM) write caches. A single-copy NVRAM cache creates a single point of failure while a dual-copy NVRAM cache is very expensive because of the high cost of NVRAM. This paper presents a new cache architecture called RAPID-Cache for Redundant, Asymmetrically Parallel, and Inexpensive Disk Cache. A typical RAPID-Cache consists of two redundant write buffers on top of a disk system. One of the buffers is a primary cache made of RAM or NVRAM and the other is a backup cache containing a two-level hierarchy: a small NVRAM buffer on top of a log disk. The small NVRAM buffer combines small write data and writes them into the log disk in large sizes. By exploiting the locality property of I/O accesses and taking advantage of well-known Log-structured File Systems, the backup cache has nearly equivalent write performance as the primary RAM cache. The read performance of the backup cache is not as critical because normal read operations are performed through the primary RAM cache and reads from the backup cache happen only during error recovery periods. The RAPID-Cache presents an asymmetric architecture with a fast-write-fast-read RAM being a primary cache and a fast-write-slow-read NVRAM-disk hierarchy being a backup cache. The asymmetrically parallel architecture and an algorithm that separates actively accessed data from inactive data in the cache virtually eliminate the garbage collection overhead, which are the major problems associated with previous solutions such as Log-structured File Systems and Disk Caching Disk. The asymmetric cache allows cost-effective designs for very large write caches for high-end parallel disk systems that would otherwise have to use dual-copy, costly NVRAM caches. It also makes it possible to implement reliable write caching for low-end disk I/O systems since the RAPID-Cache makes use of inexpensive disks to perform reliable caching. Our analysis and trace-driven simulation results show that the RAPID-Cache has significant reliability/cost advantages over conventional single NVRAM write caches and has great cost advantages over dual-copy NVRAM caches. The RAPID-Cache architecture opens a new dimension for disk system designers to exercise trade-offs among performance, reliability, and cost.

A Framework for Simple Sorting Algorithms on Parallel Disk Systems

In this paper we present a simple parallel sorting algorithm and illustrate its application in general sorting, disk sorting, and hypercube sorting. The algorithm (called the (l,m) -mergesort (LMM)) is an extension of the bitonic and odd—even mergesorts. Literature on parallel sorting is abundant. Many of the algorithms proposed, though being theoretically important, may not perform satisfactorily in practice owing to large constants in their time bounds. The algorithm presented in this paper has the potential of being practical. We present an application to the parallel disk sorting problem. The algorithm is asymptotically optimal (assuming that N is a polynomial in M , where N is the number of records to be sorted and M is the internal memory size). The underlying constant is very small. This algorithm performs better than the disk-striped mergesort (DSM) algorithm when the number of disks is large. Our implementation is as simple as that of DSM (requiring no fancy data structures or prefetch techniques.) As a second application, we prove that we can get a sparse enumeration sort on the hypercube that is simpler than that of the classical algorithm of Nassimi and Sahni [16]. We also show that Leighton's columnsort algorithm is a special case of LMM.

Selection Algorithms for Parallel Disk Systems

With the widening gap between processor speeds and disk access speeds, the I/O bottleneck has become critical. Parallel disk systems have been introduced to alleviate this bottleneck. In this paper we present deterministic and randomized selection algorithms for parallel disk systems. The algorithms to be presented, in addition to being asymptotically optimal, have small underlying constants in their time bounds and hence have the potential of being practical.

Minimizing stall time in single and parallel disk systems

We study integrated prefetching and caching problems following the work of Cao et al. [1995] and Kimbrel and Karlin [1996]. Cao et al. and Kimbrel and Karlin gave approximation algorithms for minimizing the total elapsed time in single and parallel disk settings. The total elapsed time is the sum of the processor stall times and the length of the request sequence to be served. We show that an optimum prefetching/caching schedule for a single disk problem can be computed in polynomial time, thereby settling an open question by Kimbrel and Karlin. For the parallel disk problem, we give an approximation algorithm for minimizing stall time. The solution uses a few extra memory blocks in cache. Stall time is an important and harder to approximate measure for this problem. All of our algorithms are based on a new approach which involves formulating the prefetching/caching problems as linear programs.

User-Level Parallel File I/O


 
 
 
 Parallel disk I/O subsystems are becoming more important in today’s large-scale parallel machines. Parallel disk systems provide a significant boost in I/O performance reducing the gap between processor and disk speeds. We describe a Unix-like file I/O user interface, implemented in a parallel file I/O subsystem on an MIMD machine, the nCUBE 2. Based on message passing, we develop parallel disk read/write algorithms to achieve higher parallelism when more disk drives are used. We use the closed queuing network model to analyze the effect of some tunable system parameters for our parallel file system. We then performed simulation experiments in order to obtain more realistic performance data, by comparing the original vendor-supplied file system with ours. The results indicate that the speedup in I/O performance is almost equal to the number of disk drives we use. Thus, our user-level parallel file I/O approach will provide scalable I/O performance.
 
 
 

Don't Be Too Clever: Routing BMMC Permutations on the MasPar MP-2

The authors implemented and measured several methods to perform BMMC permutations on the MasPar MP-2. Each method is coded in C and MPL and uses only the global router to transfer data among processors. Implementation in MPL limits the amount of control the programmer has over the global router. The methods used were a naive method; the PE method, which adapts the block BMMC algorithm for parallel disk systems to treat processor elements as independent devices; and the cluster method, which adapts the block BMMC algorithm to treat clusters of processor elements as independent devices. The authors also implemented variations of the first two methods to compute virtual-processor numbers in Gray-code order. Our results indicate that for random BMMC permutations, the best method overall is the naive method with the Gray-code variation. The PE methods are almost as good, especially when each processor element has several items to route. For some permutations, however, the naive method performs very poorly; the best method in these cases is the PE method with the Gray-code variant. Although one might expect the cluster method to be best, since it minimizes congestion at the global router ports, this method turns out to be the worst overall, and it is always much worse than the PE method. If it had been coded at a lower level than MPL, the cluster method may well have turned out to be the best in many cases.

Data partitioning and load balancing in parallel disk systems

Parallel disk systems provide opportunities for exploiting I/O parallelism in two possible ways, namely via inter-request and intra-request parallelism. In this paper, we discuss the main issues in performance tuning of such systems, namely striping and load balancing, and show their relationship to response time and throughput. We outline the main components of an intelligent, self-reliant file system that aims to optimize striping by taking into account the requirements of the applications, and performs load balancing by judicious file allocation and dynamic redistributions of the data when access patterns change. Our system uses simple but effective heuristics that incur only little overhead. We present performance experiments based on synthetic workloads and real-life traces.

Performing out-of-core FFTs on parallel disk systems

The Fast Fourier Transform (FFT) plays a key role in many areas of computational science and engineering. Although most one-dimensional FFT problems can be solved entirely in main memory, some important classes of applications require out-of-core techniques. For these, use of parallel I/O systems can improve performance considerably. This paper shows how to perform one-dimensional FFTs using a parallel disk system with independent disk accesses. We present both analytical and experimental results for performing out-of-core FFTs in two ways: using traditional virtual memory with demand paging, and using a provably asymptotically optimal algorithm for the Parallel Disk Model (PDM) of Vitter and Shriver. When run on a DEC 2100 server with a large memory and eight parallel disks, the optimal algorithm for the PDM runs up to 144.7 times faster than in-core methods under demand paging. Moreover, even including I/O costs, the normalized times for the optimal PDM algorithm are competitive, or better than, those for in-core methods even when they run entirely in memory.